Tesi etd-01142026-155145
Link copiato negli appunti
Tipo di tesi
Corso di Dottorato (D.M.226/2021)
Autore
PAGGETTI, FLAVIA
URN
etd-01142026-155145
Titolo
The myokinetic interface: disturbance-rejection strategies for robust myokinetic control
Settore scientifico disciplinare
ING-INF/06
Corso di studi
Istituto di Biorobotica - Ph.D. in Biorobotica
Relatori
relatore Prof. CIPRIANI, CHRISTIAN
Parole chiave
- Human-machine interface
- prosthesis
- magnet
Data inizio appello
14/05/2026;
Disponibilità
parziale
Riassunto analitico
The loss of a hand has a significant impact on an individual’s daily life, due to the countless tasks and activities that the human hand can perform. From basic social gestures like handshaking to more demanding functional tasks of everyday living, the loss of a hand results in a huge functional gap, often leading to emotional distress, depression and social isolation. In the past decades, bioengineers have actively worked toward the development of technologies and control strategies aimed at bridging this gap and replicating the complex dexterity of the human hand. Many advanced technological solutions have been developed and proved capable of mimicking several Degrees of Freedom (DoF) and the fine motor skills of the human hand. Nonetheless, users are often dissatisfied with prosthetic devices, as current systems fail to provide reliable and intuitive control that accurately translates their intended movements into prosthetic action.
Over the past decades, clinicians and engineers have actively worked towards the development of human-machine interfaces capable of filling this gap and restoring intuitive and physiologically appropriate control, either developing novel surgical strategies for bionic reconstruction and sensing technologies for high-resolution acquisition of biosignals for control. Most of the control strategies available so far rely on the transduction of efferent electrical signals from the brain. In contrast, the myokinetic interface monitors the physical deformation of muscles during contraction to extract control commands for the prosthetic devices. The approach involves implanting a multitude of permanent magnets into the residual muscles. The poses of the magnets are tracked using magnetic field sensors placed in the prosthetic socket and processed through a localization algorithm. The displacement of the magnets is mapped to the contractile state of the muscles to generate control signals for the prosthetic device.
The concept was first introduced in 2016 and was specifically developed for the control of upper-limb prostheses, demonstrating the feasibility of tracking multiple permanent magnets in real-time in a workspace resembling the human forearm.Within this interface, magnet placement and sensor positioning are critical factors influencing overall system performance, as they directly affect localization accuracy, the number and type of control signals available for prosthetic control, and the robustness of the system to noise. While magnets are implanted within muscle tissues, the sensors that detect the magnetic field are typically integrated in the prosthetic socket or positioned on the skin, where they may be subjected to shifts caused by external forces or changes in limb position. These relative movements between the signal sources (magnets) and the sensing system (magnetic field sensors) introduce artifacts in the estimated magnet displacement, potentially leading to unintended prosthetic activations. Consequently, optimizing magnet placement is essential to enhance the performance of the myokinetic interface, minimize surgical burden, and support the development of a technology that is both more effective and more readily accepted by users.
This work contributes to the development of the myokinetic interface by systematically investigating how implant configuration influences control signal generation and robustness to external physical disturbances and limb position effects. The first part examines how the number of implanted intramuscular magnets affects localization accuracy and robustness to external socket shifts. These findings are validated through the first in-human clinical trial of a myokinetic interface, with particular emphasis on signal characterization and functional performance.
Building on the outcomes and design guidelines derived from this clinical evaluation, the second part investigates how magnet placement within the residual limb, corrective strategies, and control signal selection influence the system’s robustness to physical disturbances, external loads, and changes in limb position.
Finally, the dissertation extends beyond use of the myokinetic interface in natively innervated muscles by evaluating the feasibility of integrating it with advanced surgical techniques, specifically Regenerative Peripheral Nerve Interfaces. This combined approach aims to enhance both the quality and quantity of available signals for control, providing a pathway toward more reliable and robust human-machine interfaces.
Over the past decades, clinicians and engineers have actively worked towards the development of human-machine interfaces capable of filling this gap and restoring intuitive and physiologically appropriate control, either developing novel surgical strategies for bionic reconstruction and sensing technologies for high-resolution acquisition of biosignals for control. Most of the control strategies available so far rely on the transduction of efferent electrical signals from the brain. In contrast, the myokinetic interface monitors the physical deformation of muscles during contraction to extract control commands for the prosthetic devices. The approach involves implanting a multitude of permanent magnets into the residual muscles. The poses of the magnets are tracked using magnetic field sensors placed in the prosthetic socket and processed through a localization algorithm. The displacement of the magnets is mapped to the contractile state of the muscles to generate control signals for the prosthetic device.
The concept was first introduced in 2016 and was specifically developed for the control of upper-limb prostheses, demonstrating the feasibility of tracking multiple permanent magnets in real-time in a workspace resembling the human forearm.Within this interface, magnet placement and sensor positioning are critical factors influencing overall system performance, as they directly affect localization accuracy, the number and type of control signals available for prosthetic control, and the robustness of the system to noise. While magnets are implanted within muscle tissues, the sensors that detect the magnetic field are typically integrated in the prosthetic socket or positioned on the skin, where they may be subjected to shifts caused by external forces or changes in limb position. These relative movements between the signal sources (magnets) and the sensing system (magnetic field sensors) introduce artifacts in the estimated magnet displacement, potentially leading to unintended prosthetic activations. Consequently, optimizing magnet placement is essential to enhance the performance of the myokinetic interface, minimize surgical burden, and support the development of a technology that is both more effective and more readily accepted by users.
This work contributes to the development of the myokinetic interface by systematically investigating how implant configuration influences control signal generation and robustness to external physical disturbances and limb position effects. The first part examines how the number of implanted intramuscular magnets affects localization accuracy and robustness to external socket shifts. These findings are validated through the first in-human clinical trial of a myokinetic interface, with particular emphasis on signal characterization and functional performance.
Building on the outcomes and design guidelines derived from this clinical evaluation, the second part investigates how magnet placement within the residual limb, corrective strategies, and control signal selection influence the system’s robustness to physical disturbances, external loads, and changes in limb position.
Finally, the dissertation extends beyond use of the myokinetic interface in natively innervated muscles by evaluating the feasibility of integrating it with advanced surgical techniques, specifically Regenerative Peripheral Nerve Interfaces. This combined approach aims to enhance both the quality and quantity of available signals for control, providing a pathway toward more reliable and robust human-machine interfaces.
File
| Nome file | Dimensione |
|---|---|
Ci sono 1 file riservati su richiesta dell'autore. |
|